Sunday, September 13, 2015

Part 1 - Virtualizing SAP - Why, When & How?

Last week I introduced this series about running SAP on VMware and this would be the first part of this series. As a technologist, I believe we should always look at both the sides of the coin while taking decisions and for me the first side is always business.

If you are not designing a technical solution keeping in mind the business benefits, then chances are that either your solution will miserably fail or it would never be adopted by the business since it does not keeps them into consideration.



WHY?

Working on the same principle, I thought I would start this series by talking about the Why and Whens before jumping into the Hows! Let us jump onto the reasons as to why we should virtualize SAP and how will it impact the Business Side and the Technology side of what we all know is one coin. 

I think it would be a marketing talk if I start talking about the features, stability and performance of vSphere as one of the driver for an organization to virtualize SAP. While these factors are a confidence builder, honestly speaking for a business it does not matter what platform they are running on, till the time the underlying infrastructure can serve the business demands without any compromise in the performance of business transactions which at the end of the are the backbone of any organization. Hence it is mostly for IT to decide on the platform while the business mostly looks at the reference case studies and the adoption rate by it's peers in the industry as a benchmark for choosing a platform.

Needless to say that vSphere has the most number of SAP workloads running on it if compared with any other hypervisor, because SAP itself runs there IT on VMware and I have not seen many, infact any customer NOT to follow the footsteps of there most critical software vendor. Now that we know that business will mostly follow references, it is important that we can get the right ones. It is easy to find references here on this link and I would encourage you to use them if required.

While the usual trend for business critical applications is to run on the most expensive hardware with no questions asked, we are seeing a rapid change in this trend in the last 5 years where customers are migrating from the expensive world of Unix platforms (Capex and Opex wise) to the commodity world of x86 platform. To top it all, Intel has done a fairly good job to ensure that there processors are becoming robust by the day to ensure higher standards of hardware availability. So while the biggest benefit to the business is saving tonnes of money, they also get the benefit of making a traditional application like SAP run on a platform which gives it the flexibility to Scale, Perform, Migrate, Upgrade, Update & Recover from disasters in a much more efficient manner than the traditional platforms.

At the end of the day SAP is just a simple 3 tier application, hence the 1st mantra is that the environment which you choose to run this application on should be simple.



WHEN?

While I can say NOW, it is only true for greenfield environments. For brownfield, the time to virtualize SAP has to the most critical decision for sure. While technology is no longer a limitation with the growing scale of physical x86 servers and VMware virtual machines, from a business point of view if you are planning to migrate from a Unix based environment, then you have to wait for certain opportunities which would make the move less painful or more rewarding. Let us quickly review all the opportunities: 

1- New SAP Deployment (Greenfield) - A scenario where you are building your first SAP module grounds up in a new SAP deployment. In this situation, virtualization has to be the first choice. Remember this includes all the tiers of your SAP application including the Oracle database which is licensed with the SAP license and hence you don't have to deal with Oracle licensing nightmares. One the other hand if you are worried about the WHY's then read above. 

2- SAP Upgrade Cycle - This is a no brainer, since this is your opportunity to explore the approved downtime from business around upgrading the SAP landscape. The SAP upgrade is where you could re-architect the SAP environment inclusive of Test, Development and Production to ensure that you can get the benefits of virtualization such as High Availability due to a distributed install.

3- Migration to 64 Bit Net-weaver - This is similar to a platform re-architecture and another great tipping point where the new 64 bit architecture can be built on the Virtual Platform of vSphere.

4- Expanding to New SAP Modules - This is similar to a new deployment, however if you have an existing SAP environment, you might NOT want to extend that for new modules, rather go for the virtual environment with the bells and whistles of VMware as this might be a lesson learnt for you to move your existing apps on top of the Virtual x86 Servers.

5- Hardware Refresh - I think this is the most obvious reason for you look at the x86 world as both the Capex and the Opex to run on the traditional hardware usually costs a bomb. 


Now that we have a good understanding of WHY and WHEN, it only makes sense to continue to the HOW part of the title!!


HOW?

Well, with this series, I will run your through a real life project which I recently completed for a customer in virtualizing their SAP Enterprise Resource Planning system, which essentially is the backbone of their organization. This includes, the TEST, Development and Production environments with the key VMware solutions for Virtualization, Operations & Disaster Recovery. Hence, the rest of the parts of these series would dive into each of the areas and numerous design decisions keeping in mind the requirements of the SAP application. While most the principles would be similar to architecting a vSphere environment for any application, with this series, I will call out the specific areas of consideration for SAP, right from sizing to implementation.

With that I will close this article and will see you with the next part of this series which would be:


Part 2 - Sizing SAP to run on VMware vSphere (coming soon...)


Till then.. Stay tuned!!

Share & Spread the Knowledge...




Thursday, September 10, 2015

SAP on VMware Series : It's all about the Design!

Some time back I tweeted about the value of doing a series around running SAP workloads on VMware and I got a positive response from the community as this is still considered to be a black art. I was lucky to learn from some of the best in the industry and I am referring to an ex-colleague and a great friend on mine : Kasim Hansia

While Kasim is the right person to write this series, he is too busy selling the Nutanix boxes now a days :-) Fortunately, in my recent projects, I had the opportunity to apply what I have learnt from Kasim overtime and I thought it would be useful for the community if I could bring all those learning's in a multi-part series. With this series I will take an example of a specific SAP module and design the upstream and downstream dependencies keeping in mind the Availability, Performance, Scalability, Manageability and Recoverability for the environment. The goal is to help the architects and customers who plan to migrate or do a greenfield deployment of SAP on x86 virtual platform to get the best of both worlds. The goodness of SAP, the features of vSphere and the stability of the vSphere platform.

I did a similar series around Architecting vSphere a couple of years back and it was well received by the readers. I will try and do a similar or better job on this. You can bookmark this page if you are interested as I will keep on adding the links to all the parts right on this article.




Tuesday, September 8, 2015

A Sneak Peak of our VMworld Session : MGT 4973 Mastering Performance Monitoring and Capacity Planning.

VMworld 2015 San Francisco Edition is wrapped up and I am back home to the Sunny Singapore. Being in the VMware arena for more than 7 years now, this was only my first VMworld and it was simply amazing. I had withdrawal symptoms once I was back as everything just became slow after those 4 days of Always On crazy rush.

While this post is to give you a sneak peak of how we did on our maiden VMworld Session : MGT 4973, I would also take this opportunity to thank all the fantastic people I met in the conference. The co-bloggers, amazing speakers, customers, my twitter buddies, vExperts, VCDXs, partners etc. Wow, it was an ocean of geeky minds.

Well, I will not dive into how was our session and how did we do in terms of session scores, attendees etc. as Iwan Rahabok has already shared the same in his post here.

While. we presented this session twice, we were lucky to provide a summary of this session on a Live vBrownBag video from the hangspace. I would end the post with this video. Once the full session is published, I will share the same in a consecutive post. Till then enjoy the short video.

Special thanks to #vBrownBag for giving us a hook:






Share and Spread the Knowledge!!



Monday, September 7, 2015

Monitoring VSAN hardware using vRealize Operations Dashboard.

A few days back I mentioned about the general availability of the vRealize Operations Management pack for vROps in my article here. While building content for VMworld, I tried my hands on the management pack and the metrics which are generated by vROps through this adapter and I was amazed at the amount of detail which you can get about the health of your VSAN cluster. 

I tasked myself to use these metrics to monitor the key areas of the VSAN solution. A failure in these key areas in my opinion can impact the availability or performance and hence monitoring these areas is a key. With this post, I will share one of the many dashboards which I have created for VSAN and I hope they will be useful for you. I might not share the entire dashboard for you to download and import, as some of the widgets in the dashboards are static and might not populate in your instances.

With this dashboard, I am monitoring 3 key layers of the VSAN solution:

a) Solid State Drives - SSD's are the heart of any hyper-converged solution. The performance tier of VSAN is dependent on the SSD's, hence we all understand the importance of monitoring them for failures of wear and tear.

b) Magnetic Disks - This essentially is the Capacity tier of VSAN and monitoring this is as important as any other hardware component of the VSAN solution.

c) Host Adapter - The host adapter is another crucial component which has to be monitored as it provides access to the performance and data tier.


Here is the dashboard which I have built:



Let me quickly summarize the dashboard for you:-

  • The first row is SSD Monitoring, where you can see a heatmap indicating the errors on a SSD in 7 different VSAN Clusters. If a SSD has an error the color & size of the column representing that SSD would turn black in color. You can select the faulty object and then click on the Resource Details icon to know more about the SSD Error and see if it requires a replacement by looking into the Recommendations from vROps.
  • The second row is for monitoring the magnetic disks. This is built with the same logic like the previous row of SSD. A black color box means a problematic disk.
  • The third row here is to monitor the adapter health. In this case I have faulty adapter and hence the same has been highlighted in black color on four VSAN Clusters.
Now if you want to create this dashboard in your environment, you would use the Text Widget for pulling up those icon images from an html page and use the heatmaps on the right to display the metrics related to those hardware objects.

Let's quickly run through the built:

Step 1 - Login to vROps.

Step 2 - Select Content -> Manage Metric Config (You must have vROps 6.0.1 or above)

Step 3 - Expand TxtWidgetContent

Step 4 - Click on the GREEN PLUS SIGN to add a new html.

Step 5 - Name the file sunny.html and paste the following code in the file and click on Save

sunny.html

"<!-- Codes by HTML.am -->

<!-- CSS Code -->
<style type="text/css" scoped>
img.GeneratedImage {
width:300px;height:225px;margin:10px;border-width:6px;border-color:#000000;border-style:solid;
}
</style>

<!-- HTML Code -->
<img src="http://icons.iconarchive.com/icons/thvg/slick-drives/512/SSD-Drive-icon.png" alt="Photo of the Remarkables mountain range in Queenstown, New Zealand." class="GeneratedImage">
"


Step 6 - Use the same steps to create the other 2 files. Here is the code and screenshot to guide you.

hdd.html

<!-- Codes by HTML.am -->

<!-- CSS Code -->
<style type="text/css" scoped>
img.GeneratedImage {
width:300px;height:225px;margin:10px;border-width:6px;border-color:#000000;border-style:solid;
}
</style>

<!-- HTML Code -->
<img src="http://pngwebicons.com/upload/hard_disk_PNG9180.png" alt="Photo of the Remarkables mountain range in Queenstown, New Zealand." class="GeneratedImage">




HOSTADAPTER.html


<!-- Codes by HTML.am -->

<!-- CSS Code -->
<style type="text/css" scoped>
img.GeneratedImage {
width:300px;height:225px;margin:10px;border-width:6px;border-color:#000000;border-style:solid;
}
</style>
<!-- HTML Code -->
<img src="http://ecx.images-amazon.com/images/I/411qssJgDvL.jpg" alt="Photo of the Remarkables mountain range in Queenstown, New Zealand." class="GeneratedImage">






NOTE : PLEASE ENSURE THAT IF YOU DO NOT HAVE INTERNET CONNECTIVITY ON VROPS YOU WOULD NEED TO HOST THE ICONS SOMEWHERE ON AN INTERNAL WEBPAGE AND POINT OT THAT AS <img src= ON ALL THE 3 HTML CODES.

Assuming you are done with the icon hosting and those three html files, here is what you need to DOWNLOAD. The file you just downloaded is the export of the dashboard and you would just need to import the same in your environment and you should have the final dashboard.

let's see the steps to import it.
  • Login to your vROps instance with an administrator account. (This could be to any instance of the cluster in case you are running a vROps Cluster).
  • Click on Content - > Dashboards -> Actions -> Click on the Import Dashboard.
  • Browse to the VSANHWDASH.json file and click on Ok.
  • This will bring the VSAN HARDWARE MONITORING Dashboard in the list of dashboards which you have.

Feel free to reach out through comments if you have any questions!!



Sunday, September 6, 2015

Guest Post : vRealize Operations Manager Metric Guide!




************Updated with vROps 6.6. Statkeys on 8th January 2018**************

I would like to thank my great friend Sunny Dua for giving me the opportunity to be his guest blogger.

I came across many requests both internally (field consultants / SE) and externally (Clients) for an extensive list of available metrics of vRealize Operations Manager in the form of sheet, so that one can easily refer to without accessing the product UI. The request was genuine, as vROps has rich library of metrics which can be leveraged for doing performance and capacity analysis and creating super-metrics, views, dashboards. I started looking through various available documents and internal sources, but couldn’t find any. Now in the same process, I found a wonderful stuff which is REST interface for vROps. Now coming from an Automation background, I really like to play with api’s and use them to automate as much as possible.


Now, let’s deep-dive into, how I have extracted these metrics.


vRealize Operations API documentation can be found at https://IPADDRESS OF vROPs/ suite-api/docs/rest/index.html. In this page, you can find all relevant information like, which API’s are available, sample request and response etc. The REST call that I made to retrieve these metric details is

That’s it. Is it the end of the Blog? Nope, because I know, you would be wondering, how did I get to this request, or how do I know which API call I need to fire, Right?

So, let me explain this. The first thing that we do, whenever we want to see any Metric, is we always select the Adapter Type first in GUI interface. So select the “getAdapterTypes”. The request now becomes.


















Now, fire the above request on web browser.  Here is what you will see.
























So, now we know that the adapter-kind key is VMWARE. So the call now becomes

Again, let’s see what we get after executing above request. 



























So, now we see all the Object Types available under VMWARE. Note down the resourceKinds key for the Object you want to retrieve metrics. Request now becomes

Execute the request and you see.






















That’s it. Here you get the complete http request for getting the statkeys:

Now, fire the above commands one by one replacing the resourceKinds with each object available under the VMWARE adapterkind. Once you have all the data, it’s quite easy to then dump them into XLS format.

Hope this consolidated sheet helps you and becomes your reference tool. I have also uploaded this on the following link and you can get a copy. However it is always good to know how I created this as you can do the same for any other AdapterKind.


- This is a Guest Post by Sourabh Shrivastava.


Updated with vROps 6.6. Statkeys on 8th January 2018